Compositions and methods for silencing KRAS

Abstract

The present invention relates to the inhibition of mutant KRAS sequences using RNA interference (RNAi). Additionally, the present invention provides lipid nanoparticle (LNP) compositions as delivery vehicles for RNAi agents and methods of administering them for therapeutic purposes.

Description

[Technical field]

[0001] Related application data This application claims priority to U.S. Provisional Patent Application No. 63 / 270,229, filed October 21, 2021, which is incorporated by reference in its entirety. [Background technology]

[0002] Ras mutations are associated with approximately 16% of human cancers. Kras is the most frequently mutated Ras isoform, accounting for 85% of all Ras-related cancers. Kras is tethered to the cell membrane via farnesylation. Kras cycles between an active GTP-bound form and an inactive GDP-bound form. Wild-type (WT) Kras is activated via the EGFR tyrosine kinase. In contrast, mutant kras are constitutively activated in a subset of tumor cells. Kras mutations are present in approximately 25% of tumors and are therefore one of the most common genetic alterations linked to cancer. Kras mutations are a frequent driver of lung, colorectal, and pancreatic cancer. KRAS drives 32% of lung cancer, 40% of colorectal cancer, and 85% to 90% of pancreatic cancer cases. G12C, G12D, G12V, G12R, and G13D are some of the most common KRAS mutations, based on the specific mutations present. Selective targeting of kras mutations is a promising strategy for cancer therapy, as it can complement the activity of EGFR tyrosine kinase inhibitors and reduce the side effects caused by targeting WT kras.

[0003] RNA interference (RNAi) is a gene regulation mechanism based on either small interfering RNA (siRNA) or microRNA (miRNA) that functions via incorporation into the RNA-induced silencing complex (RISC). miRNA mimics or artificial miRNAs are synthetic analogs of the physiological miR-miR* duplex. Both siRNA mimics and miRNA mimics are typically designed as oligoduplexes consisting of a guide strand and a passenger strand. In the cell, the passenger strand is degraded and the guide strand is retained in the RISC to seek out target sequences in the mRNA coding sequence or 5'-UTR or 3'-UTR and downregulate gene expression via mRNA degradation and / or translation blockade. siRNA-based gene silencing mechanisms typically require a high degree of sequence identity between the guide strand and the target mRNA, although some mismatches are tolerated at multiple positions. In contrast, miRNA-based mechanisms are highly dependent on a seed region (nt 2-7) that perfectly matches the mRNA target sequence. Strictly speaking, miRNA is defined as a natural non-coding RNA that is part of the human genome. However, it is possible to design artificial miRNA (amiR) that contains a seed region that is complementary to the target sequence in mRNA, and achieve gene silencing based on miRNA-like mechanism. Similarly, siRNA can be designed to target specific regions of genes based on its sequence complementarity with the guide strand. Depending on the overall degree of complementarity of the target sequence and the seed region, amiR and siRNA can have activity derived from both miRNA and siRNA mechanisms, respectively, and the overall gene silencing result reflects both types of activity.

[0004] Acunzo et al. (Non-Patent Document 1) designed amiRs containing seed regions that matched portions of the kras coding region that contained kras point mutations, and created six amiRs for each point mutation. In addition, a central bulge was introduced into the amiR sequence to create a 3-nt mismatch with the kras mRNA target to attenuate siRNA-like activity. Overall, these amiRs had seed regions that perfectly matched the kras mutant target, with an overall 3-nt mismatch to the mutant. However, this amiR contained an additional mismatch to the seed region of kras wt, resulting in a total of 4-nt mismatch to wt, resulting in selectivity for kras mutant over kras wt. However, this strategy resulted in amiRs with relatively low activity against the kras mutant target and a large variability in the selectivity of the amiR for kras mutants when tested in vitro.

[0005] Another strategy for targeting kras mutants is to design siRNA molecules against regions containing point mutations. Strategically, it may be advantageous to introduce mismatches so that the siRNA has one less mismatch against the mutant than the wild type. Papke et al. designed siRNAs with two mismatches each against G12C, G12D, and G13D, and three mismatches against kras wt. As a result, the final sequence EFTX-D1 is said to have silencing activity against all three mutants, while the silencing activity against kras WT is greatly reduced. However, when the present inventors tested this specific siRNA in cell lines, they found only relatively low gene silencing activity and generally poor kras mutant selectivity. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] PNAS 2017, 114:E4203-E4212 Summary of the Invention

[0007] Considering the limitations of existing approaches, the present inventors have created a new strategy for designing amiR and siRNA, and have identified amiR and siRNA with improved efficacy and selectivity against kras mutants over kras wild type, thereby reducing the possibility of side effects in therapeutic applications.In addition, amiR / siRNA is incorporated into lipid nanoparticles (LNPs) to enhance in vivo delivery.The present inventors have identified a new LNP composition that is particularly effective in the delivery of amiR and siRNA.

[0008] In one aspect, an artificial miRNA duplex for targeting a mutant kras is described herein. For example, in some embodiments, the miRNA duplex sequence for targeting a mutant kras comprises a guide strand sequence that follows the rule that (1) the 7th nt matches the mutant target sequence (mismatched to the WT sequence), and / or (2) the remainder of the amiR has one additional mismatch with the corresponding target sequence at either position 10 or 11. In another aspect, an siRNA duplex is provided for targeting a mutant kras. In some embodiments, the siRNA duplex for targeting a mutant kras is characterized in that (1) the target sequence is from the 2nd nt of codon 10 to the 2nd nt of codon 16, and / or (2) the guide strand sequence contains 0-1 nt mismatch with the point mutated target sequence (mismatch at position 4 with C to A substitution) and 1-2 nt mismatch with kras wt (position 4 and the point mutation site).

[0009] Furthermore, amiR / siRNA with the compositions and characteristics described herein were incorporated into lipid nanoparticles (LNPs) to enhance in vivo delivery. For RNAi therapeutics, in vivo delivery has been identified as a major limiting factor. Many approved siRNAs (five from Alnylam to date) are targeted to the liver, which has a high intrinsic uptake capacity and can be targeted via GalNAc conjugation of siRNA. However, for solid tumors, LNP-based strategies are the preferred option. LNPs have been shown to be efficient delivery vehicles for siRNA (e.g., Patisiran) and mRNA (BioNTech and Moderna COVID-19 vaccines) in the clinic and therefore may be able to be translated into the clinic. LNPs contain ionizable lipids, neutral lipids, cholesterol, and releasable PEG lipids. The choice of ionizable lipid is critical. pKa and geometric configuration are major considerations, along with biodegradability. Existing products utilize DLin-MC3-DMA (Alnylam), ALC-0315 (BioNTech), and SM-102 (Moderna) as the ionizable lipids.

[0010] A new amiR design strategy was developed by limiting the number of mismatches to kras mutant to 2 nt and to kras wt to 3 nt. In contrast, the previously published amiR strategy by Acunzo et al. had 3 nt and 4 nt mismatches to kras mutant and kras wt, respectively, resulting in suboptimal amiR activity and selectivity. This new amiR design strategy was based on a combination of experiments and software-based RNAi activity prediction. The new amiR will have both miR-like and siRNA-like activity. This is a novel strategy for designing kras mutant-selective RNAi agents.

[0011] Furthermore, a novel siRNA design strategy was developed by selecting sequences containing 0-1 nt mismatches with the kras mutant target sequence and 1-2 nt mismatches with the kras wild-type sequence, which resulted in novel siRNA designs with both high activity and kras mutant selectivity.

[0012] The above design strategy has resulted in amiRs and siRNAs that combine high gene silencing activity and high kras mutant selectivity, comparable to previously published strategies containing a larger number of mismatches. [Brief description of the drawings]

[0013] [Figure 1] Figure 1 shows the effect of amiR on expression levels of KRAS and pERK in cell lines. Cells were transfected with 50 nM amiR3 or amiR6. Protein levels were measured by Western blot. A: Levels of Kras expression. B: Levels of pERK expression. All levels were normalized to beta-actin as a housekeeping gene. [Diagram 2] Figure 1 shows the effect of amiR on expression levels of KRAS and pERK in cell lines. Cells were transfected with 100 nM amiR3 or amiR6. Protein levels were measured by Western blot. A: Levels of Kras expression. B: Levels of pEKR expression. All levels were normalized to beta-actin as a housekeeping gene. [Diagram 3] Figure 1 shows the effect of amiR on expression levels of KRAS and pERK in cell lines. Cells were transfected with 200 nM amiR3 or amiR6. Protein levels were measured by Western blot. A: Levels of Kras expression. B: Levels of pEKR expression. All levels were normalized to beta-actin as a housekeeping gene. [Figure 4]Figure 1 shows the effect of amiRs on KRAS and pERK expression levels in cell lines compared to the effect on NCI-H292 cells (Kras wt). Cells were transfected with 50 nM amiR3 or amiR6. Protein levels were measured by Western blot. A: Levels of Kras expression. B: Levels of pERK expression. All levels were normalized to beta-actin as a housekeeping gene. [Diagram 5] Figure 1 shows the effect of amiRs on KRAS and pERK expression levels in cell lines compared to the effect on NCI-H292 cells (Kras wt). Cells were transfected with 100 nM amiR3 or amiR6. Protein levels were measured by Western blot. A: Levels of Kras expression. B: Levels of pERK expression. All levels were normalized to beta-actin as a housekeeping gene. [Figure 6] Figure 1 shows the effect of amiRs on KRAS and pERK expression levels in cell lines compared to the effect on NCI-H292 cells (Kras wt). Cells were transfected with 200 nM amiR3 or amiR6. Protein levels were measured by Western blot. A: Levels of Kras expression. B: Levels of pERK expression. All levels were normalized to beta-actin as a housekeeping gene. [Figure 7] Figure 1. Effect of concentration and chemical modification on target regulation. A: Expression of kras. B: Expression of pERK. All levels were normalized to beta-actin as a housekeeping gene. [Figure 8] FIG. 1 shows the results of Western blot of AsPC-1 (KRAS G12D) cells treated with amiR. [Figure 9] FIG. 1 shows the results of Western blot of NCI-H292 (KRAS WT) cells treated with amiR. [Figure 10]FIG. 1 shows the relative expression levels of pERK and Kras from Western blot results of AsPC-1 (KRAS G12D) cells treated with amiR. [Figure 11] FIG. 1 shows the relative expression levels of pERK and Kras from Western blot results of NCI-H292 (KRAS WT) cells treated with amiR. [Figure 12] FIG. 1 shows the selectivity of amiR for G12D mutant Aspc-1 compared to WT Aspc-1 cells in downregulating pERK and Kras. [Figure 13] Figure 1 shows the relative expression of Kras in Aspc-1 cells after amiR treatment. mRNA levels were measured by qRT-PCR after treatment with three concentrations. [Figure 14] Figure 1 shows the relative expression of Kras in NCI-H292 cells after amiR treatment. mRNA levels were measured by qRT-PCR after treatment with three concentrations. [Figure 15] Figure 1 shows the relative expression of Kras in Aspc-1 and NCI-H292 cells following amiR treatment, with data from three concentrations averaged. mRNA levels were measured by qRT-PCR. [Figure 16] FIG. 1 shows the selectivity of amiR for G12D mutant Aspc-1 compared to WT Aspc-1 cells in downregulating Kras mRNA. [Figure 17] Figure 1 shows inhibition of Aspc-1 cell growth by amiR. Relative cell viability was determined after treatment with transfection agents or with amiR or siRNA loaded into LNPs. A: 10 nM, B: 20 ​​nM. [Figure 18]Figure 1 shows inhibition of growth of NCI-H292 cells by amiR. Relative cell viability was determined after treatment with transfection agents or with amiR or siRNA loaded into LNPs. A: 10 nM, B: 20 ​​nM. [Figure 19] 1 is a flow chart for the preparation of amiR- or siRNA-loaded LNPs. [Figure 20] FIG. 1 shows particle size distribution of LNPs after dialysis and then after sterilization. [Figure 21] Figure 1. Inhibition of PAN0403 cell growth by amiR or siRNA loaded LNPs. A: 50 nM, B: 100 nM. L1: DODMA-based, L2: DlinDMA-based, L3: DlinMC3DMA-based. A. amiR6-10GA, B. G12Dsi, C. Scrambled control, D. Seq-2 siRNA positive control. [Figure 22] Inhibition of BxPC-3 cell growth by amiR or siRNA-loaded LNPs. A: 50 nM, B: 100 nM. L1: DODMA-based, L2: DlinDMA-based, L3: DlinMC3DMA-based. A. amiR6-10GA, B. G12Dsi, C. Scrambled control, D. Seq-2 siRNA positive control. [Figure 23] Figure 1. Inhibition of HUVEC cell growth by amiR or siRNA loaded LNPs. A: 50 nM, B: 100 nM. L1: DODMA-based, L2: DlinDMA-based, L3: DlinMC3DMA-based. A. amiR6-10GA, B. G12Dsi, C. Scrambled control, D. Seq-2 siRNA positive control. [Figure 24] FIG. 1 shows the antitumor activity of amiR / siRNA in DODMA-based LNPs in a PAN0403 xenograft model. [Diagram 25] FIG. 13 shows the antitumor activity of amiR / siRNA in DLInMC3DMA-based LNPs in a PAN0403 xenograft model. [Figure 26] Figure 1 shows inhibition of tumor growth by amiR / siRNA in DODMA-based LNPs in a PAN0403 xenograft model.Positive values ​​indicate tumor inhibition compared to empty LNPs. [Figure 27] Figure 1 shows inhibition of tumor growth by amiR / siRNA in DlinMC3DMA-based LNPs in a PAN0403 xenograft model.Positive values ​​indicate tumor inhibition compared to empty LNPs. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] The embodiments described herein can be more easily understood by referring to the following detailed description and examples and the description before and after them. However, the elements, devices, and methods described herein are not limited to the specific embodiments presented in the detailed description and drawings. It should be recognized that these embodiments are merely illustrative of the principles of the invention. Numerous modifications and applications will be readily recognized by those skilled in the art without departing from the spirit and scope of the invention.

[0015] The following non-limiting examples provide further disclosure of the details disclosed in the Summary above.

[0016] First, we considered the data of past amiR design. Acunzo et al. (2017) disclosed a strategy for amiR design. Based on the western blot data of kras in this paper (Figure S3, panel C), we considered the resulting amiRs targeting the G12D mutation. Although it was claimed that these amiRs were selective for kras mutants, in fact, the western blot data showed counterselectivity to WT, low efficacy in targeting G12D kras, or induction of WT kras for some of the amiR candidates (see the data summary table below). Induction of WT kras expression was also found. This is problematic as it may increase tumor development. Further improvements in amiR design are clearly needed.

[0017] [Table 1]

[0018] Among the G12D-targeting amiR sequences in this paper, KD3 and KD6 produced some G12D-selective targeting effect and moderate WT induction. To investigate this observation, amiRs similar to KD3 and KD6 were synthesized and tested in the following G12D and WT cell lines at Bioduro-Sundia, a CRO. Based on KD3 and KD6, amiR3 and amiR6 were designed by adopting a duplex design and adding chemical modifications to increase the nuclease stability of amiR.

[0019] [Table 2]

[0020] The amiR duplexes tested were as follows and were purchased from Integrated DNA Technologies (IDT).

[0021] [Table 3]

[0022] The symbol * denotes a phosphorothioate bond. Additionally, amiR3a and amiR6a contained 2-OMe modified nucleotides, indicated by "mN."

[0023] Western blot data from Bioduro-Sundia (CRO) Transfections were performed using a commercial transfection reagent at 50 nM, 100 nM, and 200 nM. Western blots were performed to examine the effect of the various amiRs on the expression of kras and p-ERK (a downstream target of kras). The antibody used recognizes both mutant and wt kras proteins. Data from all three concentrations are shown in Figures 1-3.

[0024] Furthermore, to assess the selectivity of the amiRs, it was useful to look at the effect of the amiRs on kras and pERK in comparison to the wt control NCI-292 cell line. The results are shown in Figures 4-6.

[0025] Conclusions of this study: 1. In general, amiR6 had a stronger down-regulatory effect on kras and pERK than amiR3 in the above cells. 2. amiR3 showed a counterselectivity for G12D, whereas amiR6 had relatively good selectivity for G12D, especially in G12D homozygous Aspc-1 cells, and this effect was more pronounced for pERK than for kras, which was expressed at a relatively low level.

[0026] The effect of amiR concentration and chemical modification is illustrated in Figures 7A-B.

[0027] Conclusions of this study: 1. High concentrations of amiR appeared to upregulate kras in vitro, possibly due to the use of transfection agents containing cationic lipids and the potential for off-target activation of alternative pathways. The effect of amiR concentration on pERK was unclear. 2. Overall, the effectiveness of the different amiR overhang designs and chemical modifications tested was ranked as amiRt≧amiRa>amiR. Therefore, amiRt or amiRa was the preferred chemical structure of amiR.

[0028] To further improve target downregulation, it was then determined that the number of amiR mismatches to the kras G12D mutant sequence was reduced from 3 nt to 2 nt in the central bulge region. The newly designed amiR6 variants were evaluated by the open source DesiRm software, which generates estimates of gene silencing activity. As a result, two amiR6 variants were identified, namely amiR6-11CU and amiR6-10GA. The sequences of these amiRs are as follows:

[0029] [Table 4]

[0030] Targeting Kras mutants with siRNA It is possible to design siRNAs that match the kras mutant and contain mismatches with the kras wt sequence. In this case, the seed region of the siRNA guide strand (nt 2-7) was outside the region opposite the point mutation of the kras mutant, which distinguished this strategy from the amiR strategy described above. A recent paper (Papke et al. ACS Pharmacol. Transl. Sci. 2021, 4, 2, 703-712) reported an siRNA named EFTX-D1 that had two mismatches with G12D and three mismatches with the kras wt target sequence. It was claimed that this siRNA was able to selectively target kras G12D. However, EFTX-D1 was found to be suboptimal, as the number of mismatches was too high to sustain effective silencing of the Kras target. Therefore, improved siRNAs with fewer mismatches were designed, as shown below.

[0031] [Table 5]

[0032] The newly designed amiRs and siRNAs were evaluated in AsPC-1 cells homozygous for G12D and NCI-H292 cells homozygous for kras wt. Kras and pERK were analyzed by Western blot, while kras mRNA was measured by qRT-PCR. The selectivity for kras G12D over kras wt was calculated. The results are shown below.

[0033] Three amiR-based designs and four siRNA-based designs were synthesized by IDT and tested in G12D AsPC-1 and WT NCI-H292 cells. Western blots for kras and pERK and qRT-PCR for kras were performed at Bioduro-Sundia at concentrations of 25 nM, 50 nM, and 100 nM. The results are shown in the Western blot data in Figures 8-9.

[0034] The downregulation of pERK and kras in Aspc-1 G12D cells is summarized in FIG.

[0035] G12Dsi produced the greatest kras knockdown. amiR6-10GA and G12Dsi-4CA were highly potent in downregulating both pERK and kras.

[0036] In NCI H292 kras wt cells, the results are shown in FIG.

[0037] G12Dsi produced the greatest knockdown of kras, but not pERK. amiR6-11CU, amiR10GA, and EFTX-D1 also produced significant knockdown effects.

[0038] The selectivity for G12D over WT for various amiR and siRNA constructs is shown in FIG.

[0039] Selectivity for G12D appeared to be highly concentration-dependent and differed between pERK and kras.

[0040] At 25 nM, G12si4CA and amiR10GA had the highest kras and pERK selectivity (much better than EFTX-D1).

[0041] At 50 nM, amiR6T, amiR6-10GA, sikras14, and G12Dsi-4CA showed kras selectivity, and G12Dsi showed good pERK selectivity.

[0042] At 100 nM, G12Dsi-4CA showed good kras selectivity and some pERK selectivity. All amiRs and G12Dsi showed good pERK selectivity.

[0043] Looking at the effects across all concentrations: amiR-6T, amiR-6-11CU, and amiR-6-10GA showed preferential downregulation of KRas and pERK proteins in mutant (G12D) cells compared to KRas (WT) cells, with amiR-6-10GA and amiR-6-11CU exhibiting a stronger effect than amiR-6T. · G12Dsi knocked down KRas in both cell lines, but pERK was only depleted in the mutant cells. · G12D-4CA showed promise in reducing both Kras and pERK levels in mutant cells. Overall, G12Dsi-4CA and amiR6-10GA were the most selective for mutant (G12D) over WT KRas. These two artificial siRNAs / miRNAs are promising therapeutic candidates for KRasG12D, surpassing the previously reported EFTX-D1 (Silencing of Oncogenic KRAS by Mutant-Selective Small Interfering RNA. Papke B et al. ACS Pharmacol Transl Sci. 2021 Feb 4;4(2):703-712) in terms of efficacy and selectivity for both KRas and pERK.

[0044] Overall, G12Dsi-4CA and amiR6-10GA had the highest selectivity for G12D over WT, and amiR6T and G12Dsi also had significant selectivity for G12D.

[0045] Conclusions based on WB data Based on the efficacy and selectivity data, we deemed it reasonable to further evaluate amiR6-10GA and G12Dsi-4CA as potential G12D-selective amiR / siRNA therapeutics. G12Dsi also appeared promising due to its high kras knockdown potential and excellent pERK selectivity in G12D mutant cell lines. These constructs were more effective than the previously reported EFTX-D1 in terms of both efficacy and selectivity for kras and pERK.

[0046] Downregulation of kras mRNA by amiR / siRNA assessed by qRT-PCR Introduction Using qRT-PCR, we directly measured downregulation of kras mRNA targets. Note that amiR and siRNA have both mRNA downregulation and translational block. amiR likely has a greater effect on translation due to its mechanism. Therefore, qRT-PCR shows a greater effect of siRNA than amiR due to this difference.

[0047] Raw Data [Table 6]

[0048] [Table 7]

[0049] Additional data is shown in FIG.

[0050] [Table 8]

[0051] [Table 9]

[0052] Additional data is shown in FIG.

[0053] Gene silencing data was averaged and the targeting ratio of G12D vs. WT was calculated, as shown in FIG. 15 and FIG.

[0054] The conclusions of this study are based on the average qRTPCR data in cells (WT and mutant) transfected with 25 nM, 50 nM, and 100 nM amiR / siRNA.

[0055] Based on G12D mRNA knockdown, the hierarchy of activity was G12Dsi-4CA>>G12Dsi>amiR6-11CU>amiR6-10GA=EFTX-D1.

[0056] Based on the selectivity for G12D versus kras wt, the ranking was G12Dsi-4CA>>amiR6-11CU>amiR6-10GA=amiR6T=G12Dsi.

[0057] Therefore, the top candidates for efficacy and selectivity of target downregulation, based on qRT-PCR, were G12Dsi-4CA, amiR6-11CU, G12Dsi, and amiR6-10GA.

[0058] Overall, the conclusion of this study is that G12Dsi-4CA, amiR6-10GA, amiR6-11CU, and G12Dsi are all improved designs compared to the previously reported amiR6 (Acunzo et al.) and siRNAEFTX-D1 (Papke et al.). In fact, EFTX-D1 performed less well in terms of both efficiency and G12D selectivity. G12Dsi-4CA had the best overall G12D selectivity profile among the sequences tested, while the others are also very promising.

[0059] General approach to amiR and siRNA design : The above-described methods for designing amiRs and siRNAs can be readily applied to other kras mutants and to point mutations in general. A generalized design approach is as follows.

[0060] For amiR design, the 7th nt of the guide strand should match the mutant target sequence (which is a mismatch to the WT sequence). The remainder of the amiR should perfectly match the corresponding target sequence, except for positions 10 or 11, where an additional mismatch is introduced (e.g., using a G to A or C to U substitution). In this way, the total number of mismatches for the mutant will be 1 nt (the one in the center) and 2 nt for the WT (1 in the seed region and 1 in the center). Below is an example of this design method applied to targeting G12S (only the guide strand is listed):

[0061] [Table 10]

[0062] The same approach can be used to design amiR sequences that target G12V, G13D, G12C, and any other kras point mutation variants. In a typical amiR design, 2-3 phosphorothioate bonds are introduced at both the 5' and 3' ends. The passenger strand is perfectly complementary to the guide strand. In addition, dTdT may be added to each strand to create a 2 nt 3' overhang.

[0063] Regarding the design of siRNAs to target kras mutants, variants of G12Dsi and G12Dsi4CA can be easily designed to target other kras mutants. These variants will have 0-1 nt mismatches with the point mutation target sequence and 1-2 nt mismatches with kras wt. For example, to target kras G12C, the following guide strands can be used:

[0064] [Table 11]

[0065] The same approach can be used to design amiR sequences that target G12V, G13D, G12C, and any other kras point mutation variants. In a typical siRNA design, 2-3 phosphorothioate bonds are introduced at both the 5' and 3' ends. The passenger strand is perfectly complementary to the guide strand. In addition, dTdT may be added to each strand to create a 2 nt overhang at the 3' end.

[0066] In vitro evaluation of LNP-kras amiR / siRNA Tumor cell inhibition by 10GA, 4CA, and G12Dsi was analyzed by cell viability assays at 10 nM and 20 nM in ASPC-1 (G12D) and NCI-H292 (WT) cells using lipofectamine transfection agent, along with EFTX-D1, amiRscr, and a positive control (seq 2 from Kopke paper). 10GA, 4CA, and G12Dsi were also tested in an LNP format designed for in vivo delivery. The results are shown in Figures 17-18.

[0067] First, the LNPs generally had low in vitro activity (they were designed for in vivo use). 10GA at 20 nM appeared to have a slight effect on both cell types. This effect was stronger in H292 (WT) cells. This may be due to the stronger kras dependency of H292.

[0068] For all amiR / siRNAs using lipofectamine transfection agent, the cytotoxic effects in Aspc-1 cells followed the order 11CU = 10GA > 6T > G12Dsi > sikras14 = 4CA = EFTX-D1 = positive control (seq 2) > amiRscr.

[0069] For all amiR / siRNAs using lipofectamine transfection agent, the cytotoxic effects in H292 cells followed the order 11CU = 10GA > 6T >> G12Dsi = sikras14 = 4CA = EFTX-D1 = positive control (seq 2) = amiRscr.

[0070] Therefore, based on in vitro cell inhibition, 11CU and 10GA had the highest activity, which was higher than that of siRNA-based agents. The literature-derived sequence (EFTX-D1, seq 2) was not effective in terms of cell inhibition. amiR is a better RNAi agent in terms of cytotoxicity.

[0071] Synthesis and characterization of novel ionizable lipid-containing ethanol-free lipid nanoparticles (LNPs) The amiR and siRNA samples were synthesized by WuxiSTA according to our design and purified by HPLC.

[0072] material and method 1,2-Dioleyloxy-3-dimethylaminopropane (DODMA), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE), cholesterol, acetic acid, and sodium acetate were purchased from Sigma. 1,2-Dilinoleyloxy-n,n-dimethyl-3-aminopropane (DLin-DMA) and (6Z,9Z,28Z,31Z)-heptatriaconta-6,9,28,31-tetraen-19-yl 4-(dimethylamino)butanoate (Dlin-MC3-DMA) were obtained from MedChemExpress and Nanosoft polymers, respectively. L-histidine was purchased from Roth. Ethanol, RNAse-free water, and Slide-A-Lyzer™ dialysis cassettes (10 kD, 12–30 mL) were purchased from VWR. RNA was purchased from Wuxi AppTec.

[0073] [Table 12]

[0074] Preparation of LNPs Stock solutions of lipids were prepared in ethanol. DODMA chloroform solution (Sigma, 890899C-100MG) was evaporated at 40°C using a vacuum evaporator (Genevac EZ-2 Elite, automated program for low boiling point solvents) and then dissolved at 20 g / L in ethanol at 40°C. DLin-DMA, DLin-MC3-DMA, DOPE, and PEG were dissolved in ethanol at 40°C. 2000 -DMG was dissolved in ethanol at 20 g / L, while a cholesterol solution was prepared at 10 g / L.

[0075] Ionizable lipid / DOPE / cholesterol / PEG in a ratio of 46:26:26:2 mol / mol 2000Solutions composed of -DMG were prepared at 8 g / L in ethanol for each ionizable lipid (DODMA, DLin-DMA, and DLin-MC3-DMA) and heated to 40°C.

[0076] In parallel, an acetate buffer with 45 mM acetic acid and 5 mM sodium acetate in RNAse-free water was prepared. A 0.8 g / L RNA solution and 20% sucrose were prepared in RNAse-free water. All solutions were heated to 40°C.

[0077] Ionizable lipid / DOPE / Cholesterol / PEG 2000 The -DMG lipid solution (2.5 mL, 8 g / L) was rapidly injected into acetate buffer (2.5 mL) at 40° C. under magnetic stirring at 200 rpm. The RNA solution (5 mL, 0.8 g / L) was rapidly added to the above mixture at 40° C. under magnetic stirring at 100 rpm, followed by rapid injection of 20% sucrose (10 mL) under the same conditions. The resulting LNPs (20 mL) were subsequently dialyzed twice against 1 L of 10% sucrose and 10 mM L-histidine (pH=7.4) using a Slide-A-Lyzer cassette to remove ethanol (bath solution change after 4 h or overnight). Finally, the LNPs were filtered through a 0.45 μm PES sterile filter for sterilization and stored at −20° C.

[0078] Characterization of LNPs DLS DLS measurements were performed at 25 °C with a scattering angle of 174.8 ° on a Zetasizer Pro (Malvern Panalytical) equipped with a He-Ne laser (633 nm). The software used was ZS Explorer. A low volume plastic cell with a path length of 10 mm was filled with 70 μL of sample. The viscosity of the dispersant was corrected according to the solvent or solvent mixture used. Data were acquired in three different measurements with automatic optimization of the number and duration of runs per measurement. Results are expressed as the average of these measurements. The D of the objects his the average intensity of each population. The PDI is calculated from the autocorrelation function using the cumulant method.

[0079] Zeta potential Zeta potential measurements were performed at 25°C with a Zetasizer Pro (Malvern Panalytical) equipped with a He-Ne laser (633 nm) at a scattering angle of 174.8°. The software used was ZS Explorer. A folded capillary cell (DTS1070) was filled with 1 mL of sample diluted 1:100 in water. Data were acquired in five different automated measurements. Results are expressed as the average of these five measurements.

[0080] LNP preparation was performed at PMC Isochem, France, according to Figure 19 as follows. The resulting LNPs were characterized for particle size by dynamic light scattering. The average particle size was found to be <150 nm.

[0081] Particle size was measured after dialysis and again after sterilization, and the results are plotted in Figure 20.

[0082] The resulting LNP product has the following composition:

[0083] [Table 13]

[0084] [Table 14]

[0085] The LNPs were then evaluated for inhibition of tumor cells in vitro at the CRO Bioduro-Sandia. The protocol used was as follows.

[0086] Study design for LNP transfection and cell line activity assay 1. Purpose of the test: The effect of LNPs on the viability and proliferation of cell lines is detected.

[0087] 2. Study Design: LNP-siRNA is transfected into Bxpc-3 and HUVEC cell lines to knock down the mRNA levels of KRAS gene, and the viability and proliferation of the cell lines are detected at 24, 48, 72, 96, and 120 hours using the CTG method.

[0088] 3. Materials and Methods: 3.1 Cell lines

[0089] [Table 15]

[0090] 3.2 Reagents CellTiter-Glo® Luminescent Cell Viability Assay (Promega, Catalog No. G7573). RPMI1640 medium (Gibco, Catalog No. 11415-064) Trypsin-EDTA (0.25%) (STEMCELL, catalog number 09701). FBS (ExCell Bio, Catalog No. FND500) Phosphate Buffered Saline (PBS) (Gibco, Cat. No. C20012500BT) Penicillin / Streptomycin (100x) (Gibico, Catalog No. 15140-122) Sodium pyruvate (100 mM) (Gibco, catalog number 11360-070) Lipofectamine RNAi MAX (Thermo Fisher, catalog number 13778075) Dimethyl sulfoxide (DMSO) 100 mL (Sigma, Catalog number D2650-100 mL) HUVEC complete medium (Pricella, Cat. No. CM-0122) The table below lists the 15 LNPs.

[0091] [Table 16]

[0092] 3.3 Equipment Cell counter: Counter star (Ruiyu-biotech) CO2 cell incubator: MCO-15AC (Thermo Fisher) Pipette: BioHit Multichannel, 50~1200μL (RAININ Multichannel) Pipettes: 0.2-10 μL, 10-300 μL, 5-50 μL (Eppendorf) Centrifugation: Centrifuge ST 40R (Thermo Fisher) Water system: Milli-Q Reference system (Millipore) Perkin Elmer Envision 2104 Multilabel Reader(No. 01-094-0002)

[0093] 4. Assay Protocol 4.1 Preparation of cell assay plates: Day 1 1) Pre-warm trypsin-EDTA (0.25%) and cell medium in a 37 °C water bath. 2) Observe the cells under a microscope to assess confluence and to ensure the absence of bacterial and fungal contamination. 3) Remove the medium and wash the cells twice with 10 mL of PBS. Add 2 mL of 0.25% Trypsin / EDTA reagent to the T-75 flask. Place the flask in the incubator for a few minutes or until the cells detach. Add 7 mL of fresh cell culture medium containing 10% FBS, rinse the cells, and transfer to a centrifuge tube. 4) Centrifuge the collected cells at room temperature at 200 g for 5 minutes. 5) After centrifugation, discard the supernatant and resuspend the cell pellet in 5 mL of complete cell culture medium. 6) Take 20 μL of resuspended cells for cell counting. Using a Cell Counter Star, count the cells by adding 20 μL of cell suspension to 20 μL of dye and record the viable cell count and viability on the cell tracking sheet. 7) Adjust the volume of the suspension with complete cell culture medium to achieve cell concentration. 8) Seed cell lines at a density of 2000 cells in 90 μL per well in a 96-well plate. Incubate cell lines overnight at 37°C with 5% CO2. Perform in triplicate. Seed 10 plates for each cell line.

[0094] 4.2 Transfection of LNPs: Day 2 Regarding transfection of LNPs: Prior to transfection, the LNPs are gently mixed with the tip. Prepare a 10x concentration of LNP as follows:

[0095] [Table 17]

[0096] For 100 nM LNP, add 10 uL of 1 uM LNP to 90 uL of cells according to the plate map. Perform in duplicate.

[0097] For 50 nM LNP, add 5 uL of 1 uM LNP + 5 uL of media to 90 uL of cells according to plate map. Perform in duplicate.

[0098] There will be a total of 12 plates, each plate for one concentration and one time point and one cell line.

[0099] 4.2 Detection of CTG: Days 3-7 (24 h, 48 h, 72 h, 96 h, and 120 h) 1) Incubate the plate at room temperature in the dark for 30 minutes. 2) Thaw 3 vials of Cell Titer-Glo® Reagent at room temperature and equilibrate to room temperature before use. Protect from light. 3) Add 100 μL / well of Cell Titer-Glo® Reagent to each well. Protect from light. 4) Mix contents on an orbital shaker for 2 minutes. 5) Incubate the plate at room temperature for 10 minutes to stabilize the luminescent signal. Read the plate on the Envision.

[0100] 5. Analysis of Results Viability (%)=((luminescence of test subject−luminescence of blank control) / (luminescence of vehicle control−luminescence of blank control))×100%.

[0101] First, data for PAN0403 cells is shown in Figures 21A and 21B, followed by data for BxP3 cells in Figures 22A and 22B.

[0102] Figure 23 shows inhibition of HUVEC cell growth by amiR or siRNA loaded LNPs. A: 50 nM, B: 100 nM. L1: DODMA-based, L2: DlinDMA-based, L3: DlinMC3DMA-based. A. amiR6-10GA, B. G12Dsi, C. Scrambled control, D. Seq-2 siRNA positive control.

[0103] The results showed that amiR 10GA and amiR 12Dsi could inhibit the growth of tumor cells and produced weak inhibition in normal endothelial HUVEC cells. In contrast, scr control and seq2 siRNA showed much less selectivity to tumor cells and caused relatively greater cytotoxicity in HUVEC cells. BxPC-3 is kras wt. HUVEC is kras wt, a non-cancerous normal human vascular endothelial cell. Therefore, the lack of toxicity in HUVEC cells and BxPC-3 indicates the possibility of reduced toxicity in normal tissues.

[0104] The LNPs were then evaluated in an in vivo efficacy study at Bioduro-Sandia.

[0105] Protocol for in vivo antitumor efficacy testing of test agents in the Panc0403 subcutaneous model in B-NDG mice 1. Purpose of the study: To evaluate the antitumor efficacy of the test agents in the Panc0403 subcutaneous model of B-NDG mice.

[0106] 2. Study Design: Treatment Groups and Doses:

[0107] [Table 18]

[0108] [Table 19]

[0109] 3.Material: 3.1 Animals and housing conditions Species:Mus Musculus Strain: B-NDG mouse Age: 6-8 weeks Gender:Female Number of animals: 135 Animal supplier: Beijing Biocytogen Co., Ltd.

[0110] 3.2 Specimen Supplier: Nanothera Bioscience Storage conditions: -80℃ Product Identification:

[0111] [Table 20]

[0112] [Table 21]

[0113] Provide sufficient volume (0.157 siRNA in approximately 255 ul LNP) to reach 2 mg / kg for each dose administration. Open 1 vial at a time for 5 vials and store in 4° C. refrigerator once thawed without refrozen until used.

[0114] 4. Experimental methods and procedures: 4.1 Cell culture Panc0403 tumor cells are cultured in 1640 medium supplemented with 15% heat-inactivated fetal bovine serum, 10 ug / ml insulin, 100 U / ml penicillin, and 100 μg / ml streptomycin at 37°C in an atmosphere of 5% CO2 in air. The tumor cells are routinely passaged 2-3 times per week. Cells growing in logarithmic growth phase are harvested and counted for tumor inoculation.

[0115] 4.2 Tumor inoculation and grouping Each mouse was inoculated with PAN0403 tumor cells (5 × 10 per mouse) in 0.1 mL of RPMI 1640 medium containing 50% Matrigel for tumor development. 6 The tumor volume (approximately 125 mm 3 90 animals will be randomized using block randomization in Excel based on the randomization strategy (SAR) for each group. This will ensure that all groups are comparable at baseline.

[0116] 4.3 Observations All procedures related to the handling, care, and treatment of animals in this study will be performed in accordance with guidelines approved by BioDuro's Institutional Animal Care and Use Committee (IACUC), which follows the guidance of the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC). During routine monitoring, animals will be checked for adverse effects of tumor growth and / or treatment on normal behavior, such as mobility, food and water consumption (by observation only), and weight gain / loss (weight measured twice weekly during the pre-dosing period and daily during the dosing period and recorded twice weekly), dulling of eyes / coat, and other abnormal effects including tumor ulceration. Unexpected deaths and observed clinical signs will be recorded based on the number of animals in each subset. Animals will not be allowed to become moribund.

[0117] 4.4 Tumor Measurement Tumor volumes were measured twice weekly in two dimensions using calipers and calculated according to the formula V = 0.5a × b 2 Use to calculate the volume in mm 3 where a and b are the long and short diameters of the tumor, respectively.

[0118] The data obtained are illustrated in Figures 24-25. The tumor growth inhibition (TGI) values ​​are shown in Figures 26-27.

[0119] The data showed that the LNP formulation had a significant effect on inhibiting tumor growth.

[0120] DODMA-based LNPs showed that the best performing amiR6-10GA produced a TGI of 15% over vehicle control (p<0.01). In contrast, Seq2 siRNA was shown to promote tumor growth (TGI=-28%). This indicates that both amiR6-10GA and amiR6-12Dsi are superior to Seq2, a nonselective siRNA reported in the literature.

[0121] The DLinMC3DMA / DOPE / PEG2000-DMG formulation resulted in a TGI of greater than 33% for both amiR6-10GA and amiR6-G12Dsi treatments compared to scrambled control-loaded LNPs (p=0.0001 for both agents).

[0122] The TGI can be further improved by further increasing the dosage by increasing the concentration of amiR / siRNA-LNP. Concentration of LNP can be easily achieved by tangential flow diafiltration (TFF), which has already been successfully performed in the laboratory. The above amiR and siRNA-loaded LNPs can be further combined with other agents, such as chemotherapeutic agents, kinase inhibitors, angiogenesis inhibitors, and immune checkpoint inhibitors, to achieve even higher TGI values.

[0123] Further analysis of results For DlinMC3DMA-based LNPs, the targeted APIs (amiR6-10GA and amiR6-G12Dsi) were better than the controls (empty LNP and LNP-scrambled control), and the improvement was greater in vivo than in vitro.

[0124] DlinDMA-based LNPs were not selected for in vivo testing due to poor in vitro results.

[0125] In vitro, the following observations have been made: 1. Mutant / WT selectivity is significantly higher for targeted APIs (e.g., amiR6-10GA and siRNA-12Dsi) compared to scrambled controls (as examined by biomarker and viability studies in Panc0403 vs. BxPC-3 cell lines). seq2 siRNA was found to be cytotoxic to HUVEC cells. As the MTD increases, this means there is a great potential to further improve the TGI with newly designed amiRs and siRNAs to improve by dose escalation. 2. The scrambled control (transfected, no LNP) is much less effective at downregulating the biomarker compared to the targeted API. 3. The scrambled control (transfected, no LNP) is significantly less effective than the targeted API. 4. The mutant-specific siRNA (EFTX) reported in the literature is less selective than our targeted API (transfected, no LNP).

[0126] [Table 22]

[0127] All nucleotides are RNA, * denotes a phosphorothioate linkage, and mN denotes a 2'-O-methyl substituted nucleotide.

Claims

1. An artificial miRNA duplex for targeting a mutant of kras, the guide strand sequence of which follows the following rule: (1) the seventh nucleotide matches the mutant target sequence (which is a mismatch to the WT sequence); (2) the remainder of the guide strand sequence has one additional mismatch with the corresponding target sequence at either position 10 or 11. According to, artificial miRNA duplex.

2. An artificial miRNA duplex comprising the sequence of SEQ ID NO: 8 or 9 for selectively targeting KRAS G12D.

3. The artificial miRNA duplex of claim 1 , wherein the miRNA duplex contains chemical modifications to enhance its nuclease stability and gene silencing efficacy.

4. An siRNA duplex for targeting a mutant kras, the target sequence of which is from the second nucleotide of codon 10 to the second nucleotide of codon 16, and the guide strand sequence contains 0-1 nucleotide mismatch with the point mutated target sequence (a mismatch at position 4 with a C to A substitution) and 1-2 nucleotide mismatches with kras wt (at position 4 and the point mutation site).

5. 1. An siRNA duplex comprising the sequence of SEQ ID NO: 11 or 12 for selectively targeting KRAS G12D.

6. The siRNA duplex of claim 4, wherein the siRNA duplex contains chemical modifications to enhance nuclease stability and gene silencing efficacy.

7. A pharmaceutical composition comprising the artificial miRNA duplex of any one of claims 1 to 3 or the siRNA duplex of any one of claims 4 to 6 for treating neoplastic diseases associated with kras point mutations.

8. The pharmaceutical composition described in claim 7, wherein the targeted kras point mutation is G12C, G12S, G12V, or G13D.

9. The artificial miRNA duplex described in claim 1, wherein the targeted kras mutation is G12C, G12S, G12V, or G13D.

10. The siRNA duplex of claim 4, wherein the targeted kras mutation is G12C, G12S, G12V, or G13D.

11. 8. The pharmaceutical composition of claim 7, wherein the neoplastic disease is selected from pancreatic cancer, lung cancer, and colorectal cancer.

12. A composition comprising an artificial miRNA duplex described in any one of claims 1 to 3 or an siRNA duplex described in any one of claims 4 to 6, wherein delivery of the artificial miRNA duplex or siRNA duplex is achieved via incorporation into nanoparticles.

13. The composition of claim 12, wherein the nanoparticles are lipid nanoparticles containing one type of ionizable lipid and a releasable PEG-modified lipid.

14. The lipid nanoparticles may comprise an ionizable lipid / DOPE / cholesterol / PEG 2000 14. The composition of claim 13, wherein the ionizable lipid comprises DODMA, DLin-DMA, and DLin-MC3-DMA in a ratio of 46:26:26:

2.

15. A pharmaceutical composition comprising an artificial miRNA duplex described in any one of claims 1 to 3 or an siRNA duplex described in any one of claims 4 to 6, and a pharmaceutically acceptable carrier.

16. A pharmaceutical composition described in claim 15 for treating cancer with a KRAS mutation.

Images